this post was submitted on 25 Jun 2026
795 points (99.3% liked)
Technology
86333 readers
2979 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If you just throw every task at it, I can completely imagine that it'll cost you more than any gain in efficiency. To some degree I think this is true of some coding tasks (not all, most are pretty efficient pattern matches, and that's what it does well).
But using LLMs to build tools and pipelines that stand alone (no AU built into it) and enable human productivity seems like a far higher leverage use.
The programming version is building the libraries and abstractions that are robust and well tested, so that regular developers can quickly build and refine the features.
Or building the reporting dashboard. Or whatever.
The cost is only going to go up, and the companies that lock in some non-AI process improvements before the hike will likely be smiling
Especially because a statistical model of language has very limited valid usecases. Many tasks that people use LLM for do not make any sense and cannot be accomplished by a statistical model of language. It can only output a statistical approximate of what a solution to the task might look like, not an actual solution.
It constructed an entire feature-full library for a UV sensor that just came on the market, by me just throwing the PDF datasheet at it and saying "make library pls".
It's the actual solution. Maybe the statistical approximate just coincidentally lined up with the actual solution? Either way, works for me. Second time it did that, too. Worked for a LiDAR sensor earlier.
This is almost the perfect task for them. I think if them as pattern matches. They have patterns for libraries in their training, and you gave it a technical spec, and it pattern matched it across to a library.
On top of that, you can verify it and reuse it. But regenerating it every time wouldn't be a good use both for the cost, and the risk of subtle issues that don't get noticed. Same argument as for any library.
Probably because I've been doing this so long, I often find it easier and more precise to describe things in code or pseudo code than common English, which often aides my use of LLMs.
On another note, I'm curious what you're making. My "when I get time" project is to use an old STM32 drone flight controller to do some basic robotics, which will only be possible because I think the LLM will pattern match me out of trouble getting an embedded C program compiling after 20 years out of that game.